CSS testing steps

There are a few basic testing steps provided with Hardy that cover a lot of use cases. For complete test coverage, you will also need to write your own custom steps but we'll look at the defaults first.

Tests

Then "<css selector>" should have "<css property>" of "<value>"

Testing for equality in CSS values. When the property measured is a colour, the values (measured and expected) are both converted to RGBA to ensure consistency. There is also a small amount of variation allowed as some browsers calculate colours slightly differently. This tolerance level is very, very small and not usually noticable.

Then "<css selector>" should have "<css property>" of "<value>" or "<value>" or ...

Same as the previous test but allows a few possible values. Useful if you want to ensure that all text element colours are from a predefined palette.

Then "<css selector>" should have "<css property>" of "<greater|less>" than "<value>"

To make tests sturdier and more reliable, it is best to not be completely exact in measurement. This step allows you to name a minimum or maximum for any numerical value. Using two of these steps together, you can create a range of acceptable values.

Then "<css selector>" should have offset "<top or left>" of "<value>"

Calculates the exact position of the element within the page and reports it back. This is measured regardless of the rules applied so that even if an element has top: auto for example, will still return the actual number of pixels from the top of the document.

Then "<css selector>" should look the same as before

This step does an image diff comparison on the element supplied. This element could be anything from a single li to the entire page.

Basic steps

These are the ones needed to open a web page and perform basic functionality. There are many more examples of these steps in the original WebDriverJS project used behind-the-scenes in Hardy.

Given I visit "<web page URL>"

Hopefully self-explanatory. Opens the specified page in the browser.

Given I wait for "<n>" seconds

Allows for interactive testing. If a particular action takes a length of time, you can pause the test for that length of time. Note: using this too often will (predictably) slow down your test runs.

Given I wait for "<selector>" to be present

Another helper for interactive testing. This will loop for a while (maximum of 5 seconds) until there is an element that matches the selector specified. If there isn't a match after 5 seconds, this fails.

Given the window size is "<width>" by "<height>"

This step allows the browser size to be controlled. This can make testing responsive designs easy.

Then I enter '<string>' into '<css selector>'

This behaves as though the user typed the string into the input identified by the selector.

Then I submit the form '<css selector>'

Equivalent to pressing submit on the form identified by the selector.

There is already quite a lot that can be accomplished with these steps but you will probably also need to write your own custom steps for complete coverage.

Beta